<libv>
hrm, serial is gone on a20 when sunxi-uart.0 is loaded... script.bin is perhaps to blame
NightShade has quit [Ping timeout: 272 seconds]
<spv>
oliv3r: in reference to the optimusboard, I believe there was a discussion about this earlier, roughly 14-16 hours ago.
<rz2k>
libv: had same with burned-out uart on a10 board, only removing the serial support helped
<rz2k>
hope you didnt go for 5v uart instead of 3v3 as I did for that board by mistake :p
Seppoz has quit [Ping timeout: 260 seconds]
vector80 has quit [Quit: Konversation terminated!]
woprr has quit [Remote host closed the connection]
<Turl>
libv: correct speed on cmdline?
<Turl>
libv: console=ttyS0,115200
<Turl>
the default is 9600 iirc
egbert has quit [Ping timeout: 246 seconds]
egbert has joined #linux-sunxi
woprr has joined #linux-sunxi
atiti has quit [Ping timeout: 246 seconds]
bsdfox has quit [Ping timeout: 252 seconds]
pacopad has quit [Ping timeout: 252 seconds]
atiti has joined #linux-sunxi
pacopad has joined #linux-sunxi
bsdfox has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 245 seconds]
pacopad has joined #linux-sunxi
spv has quit [Remote host closed the connection]
pacopad_ has quit [Ping timeout: 252 seconds]
pacopad has quit [Ping timeout: 260 seconds]
pacopad has joined #linux-sunxi
rz2k has quit []
wolfy has quit [Ping timeout: 252 seconds]
pacopad has quit [Ping timeout: 246 seconds]
pacopad has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
<libv>
Turl: bang on :)
<libv>
Turl: i think you told me that once before, about a year or so ago :)
<Turl>
libv: may have been :p
<Turl>
libv: go wiki it ;)
<libv>
just opened up the uart page :)
pacopad has joined #linux-sunxi
spv has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad_ has joined #linux-sunxi
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
pacopad_ has quit [Ping timeout: 272 seconds]
spv has quit [Quit: Ex-Chat]
<libv>
done :)
<libv>
tomorrow i will fill out the final sections of the new device howto
spv has joined #linux-sunxi
<libv>
and send in the two patches for u-boot and sunxi-boards for this new tablet
pacopad has joined #linux-sunxi
pacopad has quit [Ping timeout: 260 seconds]
pacopad has joined #linux-sunxi
pacopad has quit [Ping timeout: 252 seconds]
pacopad has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
tkoskine_ has quit [Ping timeout: 246 seconds]
spv has quit [Quit: Ex-Chat]
pacopad has quit [Ping timeout: 245 seconds]
pacopad_ is now known as pacopad
spv has joined #linux-sunxi
tkoskine has joined #linux-sunxi
pacopad has quit [Ping timeout: 245 seconds]
pacopad has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
tkoskine has quit [Read error: Connection reset by peer]
tkoskine has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad_ has quit [Ping timeout: 240 seconds]
pacopad has joined #linux-sunxi
pacopad has quit [Read error: Operation timed out]
pacopad has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 240 seconds]
pacopad has joined #linux-sunxi
pacopad_ has quit [Ping timeout: 245 seconds]
tkoskine has quit [Read error: Connection reset by peer]
tkoskine_ has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 260 seconds]
pacopad_ is now known as pacopad
pacopad has quit [Read error: Operation timed out]
pacopad has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 240 seconds]
pacopad_ is now known as pacopad
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad has joined #linux-sunxi
pacopad_ has quit [Ping timeout: 246 seconds]
bsdfox has quit [Ping timeout: 245 seconds]
npcomp has quit [Quit: Changing server]
npcomp has joined #linux-sunxi
bsdfox has joined #linux-sunxi
bsdfox has quit [Changing host]
bsdfox has joined #linux-sunxi
npcomp is now known as Guest24590
Guest24590 has quit [Client Quit]
bsdfox has quit [Ping timeout: 252 seconds]
bsdfox has joined #linux-sunxi
<mrnuke>
oliv3r: I think I fixed my RAM not working at high frequencies: here's what I did:
<mrnuke>
I used a timer to see how many ticks it takes to do 1 billion sdelay loops (came to about 8 ticks per loop)
<mrnuke>
so I replaced sdelay calls with udelay, adjusting the parameters accordingly
<mrnuke>
that gives a more consistent delay, independent of the CPU clock, and seems to init RAM at 480 MHz just fine
<mrnuke>
since I made a little mistake in my calculation, that also increased the delays by a factor of 24, and I think I just wasn't waiting enough on certain steps
<mrnuke>
bad math fixed it (don't tell anyone I have a degree in math)
spv has quit [Quit: Ex-Chat]
tkoskine_ has quit [Read error: Connection reset by peer]
npcomp_ has joined #linux-sunxi
tkoskine has joined #linux-sunxi
npcomp_ has quit [Client Quit]
npcomp_ has joined #linux-sunxi
npcomp_ has quit [Quit: leaving]
npcomp_ has joined #linux-sunxi
npcomp_ is now known as npcomp
Quarx has joined #linux-sunxi
bsdfox has quit [Ping timeout: 246 seconds]
prasannapete has joined #linux-sunxi
Black_Horseman has quit [Ping timeout: 272 seconds]
<oliv3r>
mrnuke: lol
<oliv3r>
mrnuke: but that is extremly good to know
<oliv3r>
mrnuke: i mean how figity it is :)
<oliv3r>
mrnuke: i wonder if it is safer to do the delays with a timer so they are guaranteed a certain time
<mrnuke>
always do them with a timer.
<oliv3r>
mrnuke: aye, but currently we don't
<wens>
speaking of delays, allwinner usb driver does delays with no-op while loops :(
<mrnuke>
just replace sdelay with udelay. udelay is already implemented in sunxi/uboot
<oliv3r>
sweet, eva offered me an optimus board :)
<mrnuke>
(and adjust the parameters accordingly. I divided them by 8)
<mrnuke>
sdelay depends on the CPU clock, and how efficient it is at branch prediction, etc. It's almost as good as calling udelay(random()) :p
<wens>
oliv3r: nice!
<oliv3r>
wens: how would one wish a happy spring festival in english to a chinese person?
<oliv3r>
Like we would wish eachother Happy New Year
<oliv3r>
speaking of, HAPPY NEW YEAR! :p
<wens>
you mean Chinese New Year?
<oliv3r>
mrnuke: i know it was a bad design idea, but 'its what we have' :p
<mrnuke>
oliv3r: optimus board? Sweet
<oliv3r>
mrnuke: since we have plenty of timers, that go unused otherwise anyway, might aswell use them for dram init
<oliv3r>
wens: NO! (yes) Spring Festival!
<wens>
oliv3r: Mostly just Happy New Year :p
<oliv3r>
wens: I watched discoveries 'revealed: chinese spring festival' yesterday :)
<n01>
mripard: I have a problem with the irqchip for NMI. The problem is that when my ISR for the NMI returns all the interrupts are disabled and my board hungs. Any idea why?
<oliv3r>
mnemoc: the rest should be sorted more less
<n01>
specifically any idea why interrupts could not be enabled again after the ISR?
<oliv3r>
wens: yeah that is exactly what i was after
<oliv3r>
wens: though i allready replied to eva :p but i'll use that when the time is right again
<wens>
oliv3r: it's still a bit early to use it
<oliv3r>
wens: that is true
<wens>
i would use it right after Chinese New Year
<mnemoc>
oliv3r: hate cgi based apps :\
<oliv3r>
mnemoc: if you have something else; feel free :)
<mnemoc>
oliv3r: I wish I had time to code all the toys I have sketched down already :\
bsdfox has quit [Ping timeout: 260 seconds]
<oliv3r>
mnemoc: lol
<hramrach>
heheh
<oliv3r>
mnemoc: time is not on our side
<oliv3r>
mnemoc: eva said she'll send me an octa as a gift; to which i replied; i thankfully await your gift accompanied with the usermanual and SDK so we can start supporting it fast
<mnemoc>
:D
<mnemoc>
oliv3r: I'll do the nginx/mailman/cgi stuff once arriving back home
<mripard>
n01: hmmm
* mnemoc
wants a octa board gitf too :'( </envy>
HeHoPMaJIeH has joined #linux-sunxi
<mripard>
n01: is it always the case?
deasy has joined #linux-sunxi
<oliv3r>
mnemoc: arriving from and when?
<mnemoc>
oliv3r: from work, this evening
<oliv3r>
mnemoc: ohw, ok no rush
<oliv3r>
mnemoc: then we should have mailman tomorrow :)
<mripard>
I mean, that means that only the first interrupt you get from the device works and then nothing happens, or is it a bit more random than that?
<oliv3r>
i'll sort the rest
bsdfox has joined #linux-sunxi
<n01>
mripard: it happens every time. GIC handler -> irqchip handler -> handle_level -> regmap handler. then when the GIC handler returns after the chain, the IRQ thread for reading i2c register of axp is called but everything at this point deadlocks
<n01>
if I make the chain shorter, let's say GIC handler -> irqchip handler, with the irqchip handler empty again when the handler returns everything is stuck
<mripard>
and if you don't use regmap?
<n01>
atm I need regmap to use the power button for IRQ
<n01>
I have no idea what could be
rellla has joined #linux-sunxi
<mripard>
why do you need regmap for that? isn't it just about factorizing accessors (with some caching) ?
hipboi has joined #linux-sunxi
AreaScout has joined #linux-sunxi
<n01>
I need regmap_add_irq_chip to register the irq with the irqchip
wingrime has joined #linux-sunxi
<n01>
I think I could avoid it using userspace tools (i2cset/get) to enable irqs on axp
<n01>
I can give it a shot but I don't thing that the problem is regmap
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
netlynx has joined #linux-sunxi
<mnemoc>
oliv3r: can we get smtp-auto/tls and forwards working first? :p
<mripard>
well, again, you don't need regmap_add_irq_chip, the only functions that you need to call are merely request_irq and irq_domain_add_*
<oliv3r>
mnemoc: pff, always the hard stuff first; did you enable submission port? (587)
<oliv3r>
but for smtp auto I do think you need a (virtual) mailbox, even if you decide to forward everything
<mnemoc>
oliv3r: fine then, but we need to handle those forwards
<oliv3r>
okay
<mripard>
n01: the IRQF_ONESHOT part is interesting
<mripard>
I wonder if it doesn't mask too much interrupts
<n01>
iirc I tried also without, no luck. I'll check once home tonight.
<n01>
mripard: btw, axp uses a low level irq type for signaling to NMI whereas GIC seems working only with rising edge / high level type.
<mnemoc>
oliv3r: also we can make a "dummy" dev@ subscribe the google groups now, to start collecting mails, and then inject the older
<mnemoc>
oliv3r: not sure if via nntp, raw copy into maildir or enabling imap
hipboi has quit [Ping timeout: 272 seconds]
<mripard>
n01: and the NMI is connected to the GIC? not directly to the CPU?
<mripard>
and iirc the GIC can work with all the triggers
hipboi has joined #linux-sunxi
<n01>
mripard: I'm assuming NMI is connected to GIC, even though it has special registers for masking/unmasking. in my implementation NMI is child of GIC.
<oliv3r>
but i'll sort the aliases, smtp auth etc now
<oliv3r>
with the cert and if that works; we can move over aliases etc
popolon has joined #linux-sunxi
popolon has joined #linux-sunxi
<oliv3r>
mnemoc: just let me sort kinds of other things first
<mripard>
n01: hmmm, that's odd
<mripard>
I've used the GIC two weeks ago to do some dual-edges triggering
<n01>
if in DTS I specify IRQ_TYPE_EDGE_FALLING or IRQ_TYPE_LEVEL_LOW my probing fails
<oliv3r>
mnemoc: did i just pm you?
<oliv3r>
i have to many windows I don't find it anymore
enrico_ has joined #linux-sunxi
<mripard>
n01: it was in a vendor BSP, so it might just be that no one contributed it yet.
<mripard>
or that they had some hack to make it work :)
<mripard>
ah, no, never mind
<mripard>
it was going through another interrupt controller, not the GIC
<n01>
:)
RzR has quit [Read error: Connection reset by peer]
RzR has joined #linux-sunxi
panda84kde has joined #linux-sunxi
e-ndy has joined #linux-sunxi
<oliv3r>
mripard: have you heard of the new DMAengine driver in the latest 3.4 code dump?
RzR has quit [Excess Flood]
<mnemoc>
oliv3r: in irssi I use a nice helper called /go
<mnemoc>
so /go ol<tab> does the job
<oliv3r>
i do /window <11+>
RzR has joined #linux-sunxi
<oliv3r>
i try to have less then 10 :)
<mnemoc>
I rarely can go below 30 :(
<mripard>
n01: hmmmm, I don't really have an idea, sorry :(
<oliv3r>
i did close my hacksung, tvh etc but i try to keep them only to the relevant ones
<oliv3r>
i should leave u-boot
<n01>
mripard: no prob. I'll investigate the IRQF_ONESHOT.
<mripard>
n01: you can try to ask maz whenever he comes back from his holyday, he might have an idea of what's going on
<n01>
ok. otherwise a guy in #armlinux suggested me to make NMI irqchip the parent of GIC. So I can forward every not-NMI IRQ to GIC
gzamboni has quit [Ping timeout: 240 seconds]
gzamboni has joined #linux-sunxi
_massi_ has joined #linux-sunxi
<wens>
0
<wingrime>
-rwxr-xr-x 1 alex alex 1036176 Окт 30 11:09 libiconv.a
<wingrime>
thats in cedar binaries
<wingrime>
they are seriosly link something with GPL
<wingrime>
LGPL
<wingrime>
oliv3r: thats not best thing in lastest 3.4
<wingrime>
oliv3r: best thing also, that all dma related drivers now use dma engine
<oliv3r>
wingrime: it is known to infringe
<oliv3r>
lkcl: ^
<wingrime>
oliv3r: also, nand are blob again
<wingrime>
oliv3r: and I see aw nand can be used using FUSE programm
<libv>
puneet's back!
<wingrime>
?
<n01>
lol
<oliv3r>
libv: oh please no
<oliv3r>
libv: what's the best time to arrive on friday? ;)
<oliv3r>
wingrime: aw nand fuse? now that! is very interesting ;)
<oliv3r>
btw, i got a company profile from eva, in their product history, AXP182 was their first chip :) axp 188, 189, 173, 192, 199 and THEN the F-series followed after that
tkoskine has quit [Ping timeout: 240 seconds]
<oliv3r>
Oh, "Open Source SUpport' is listed as one of their 'strengths'
libv has quit [Ping timeout: 260 seconds]
libv has joined #linux-sunxi
<oliv3r>
libv: wb
<libv>
:)
<libv>
did what i just stated make it through?
<oliv3r>
puneet B being back?
<n01>
oh btw I'm staying at Hotel Chambord
<oliv3r>
i now no longer wish to read my mail
<libv>
ah, it didn't
<libv>
oliv3r: for fosdem, i tend to arrive in time for food, so late afternoon early evening. I personally tend to arrive around 18:00 so i can drop off my stuff at the hotel and then run off to the chosen first location
<oliv3r>
chosen first location?
<libv>
i will be in the ibis grand place btw, which i think the fosdem organizers use for their speakers as well
<oliv3r>
i'm debating wether to arrive by 17:00 - 18:00 or eat at home and just arrive at 21:00
<oliv3r>
libv: novotel
<libv>
oh, they changed :)
<libv>
well, i'm right next door then
<libv>
i tend to go to this one place on the grand place, which is relatively honest price wise
<libv>
and they usually have tons of free space
<libv>
ah, no, they haven't changed. my memory is just bad.
<oliv3r>
lol
<oliv3r>
old man! :)
<libv>
no, alcoholic man :p
<oliv3r>
haha
<oliv3r>
omg, they even have a opensource 'sgrength' page in their marketing PDF saying, and I quote
<oliv3r>
GPLv3, Free as in Freedom
<libv>
last year, on monday, i first had breakfast, then had a 3h sleep, and then packed and left for home
<oliv3r>
hehe
<libv>
the previous 2 nights were only marginally better
<oliv3r>
well it is 'only' a 3 hour travel
tkoskine has joined #linux-sunxi
drachensun has quit [Remote host closed the connection]
lgentili has joined #linux-sunxi
lgentili has quit [Remote host closed the connection]
EGM has joined #linux-sunxi
<wingrime>
oliv3r: better ask eva about blobs in kernel
<wingrime>
oliv3r: isplib.a , nand
hipboi has quit [Ping timeout: 240 seconds]
<oliv3r>
she's marketing, she probably won't know, and lkcl asked to to communitcate with allwinner but inform him
<oliv3r>
lkcl: allwinner marketing slides lie!
<oliv3r>
lkcl: and there's blobs in the latest SDK + infrigement!
<mnemoc>
according to her linked in page eva has a degree in english literature, not on marketing, business or technically related
<wingrime>
mnemoc: so how she become part of aw?
<mnemoc>
been able to speak english fluently :p
<n01>
is she hot? :D
<wingrime>
oliv3r: at least note she about errors in PDF ))
EGM has quit [Ping timeout: 240 seconds]
<mnemoc>
oliv3r: we can use those slides as a tool *granting* they *want* to comply and these violations are just a naive mistake they tend to commit over and over
<mnemoc>
n01: no
<mnemoc>
oliv3r: and as they *want* to comply we only have to point out the mistakes and ask for the sources to fix them
<mnemoc>
oliv3r: she already loves you, so give the next step ;-)
<oliv3r>
wingrime: that I shall ;)
<oliv3r>
mnemoc: love? heh
<oliv3r>
she was excited about the book :)
EGM has joined #linux-sunxi
wolfy has joined #linux-sunxi
kz1 has quit [Ping timeout: 240 seconds]
hipboi has joined #linux-sunxi
<deasy>
aroooookuuuuuuuuuux
kz1 has joined #linux-sunxi
<oliv3r>
deasy: i don't htink he can hear you, increase the volume!
<mripard>
for now, I only set the machine model and the soc compatible
<mripard>
now comes the boring stuff.
<mnemoc>
mripard: somehow I still believe extending fexc to output a dts makes more sense, and also using soc-detect instead of trusting the machid
<deasy>
LOULOULOUlOUlOULOU ?
<mripard>
(and the bootargs)
<mripard>
deasy: dummy debug....
<mripard>
mnemoc: you want both actually.
<mripard>
fexc won't allow you to boot on any device out there
* libv
has exhausted his wiki time quota for 2014 :p
<deasy>
as arokux build so quickly the kernel for cubie i wanted to test the 3.13 for make feedback :)
<mnemoc>
fexc takes the .bin extracted from a running device
<mripard>
you need to extract the script.bin, run fexc, create the DT, reupload the dtb, boot it
<mripard>
which is pretty much what is done for now, only automatic
<mnemoc>
mripard: but babelfish could become part of sunxi-tools and share a libfex
<mripard>
and requires you to upgrade u-boot
<mripard>
here, you just have to update the kernel image, and that's it
<mripard>
and dtc can always extract the DT out of the running system afterwards
<mnemoc>
a layer like babelfish can indeed help booting from nand today by adding the missing meminfo to ATAGs too
<mripard>
and I'm already using fexc libfex
<mripard>
with some really minor modifications
<mripard>
(like include path, s/abort/return/, etc.)
<libv>
why not have both solutions compete on equal terms, documented equally on our wiki, and see which is best liked/supported in a years time?
<mripard>
I really believe we need both features, so I don't see why one should go away, but yeah, sure
<mripard>
hopefully, both will go away
<mripard>
but that's probably a wishful thinking
<libv>
mripard: that'll take several further allwinner generations
<mripard>
yep, true
<libv>
and we will never support all devices that have shipped already, some (by then) ancient ones might turn up in many years time
<libv>
is anyone using an rtl8188eu on sunxi-3.4
<libv>
i am unable to modprobe it, FATAL: Error inserting 8188eu (/lib/modules/3.4.75/kernel/drivers/net/wireless/rtl8188eu/8188eu.ko): Cannot allocate memory
Black_Horseman has joined #linux-sunxi
<libv>
i'm first going to set up gadget before i will try to debug that one
<mnemoc>
mripard: can you submit the changes for libfex? :)
kayhan has joined #linux-sunxi
kayhan has left #linux-sunxi [#linux-sunxi]
ganbold_ has joined #linux-sunxi
<mnemoc>
fel should become a library too
<libv>
aha, the fex file seems off.
<oliv3r>
fexc should have a -t option (test) :p
loulouxiv has joined #linux-sunxi
<oliv3r>
with input being .fex .bin
<mnemoc>
refactoring a library out and adding a decent configure/Makefile.am thing first, would clean the things enough to be able to start improving them
<libv>
ah, and g_ether is broken on a20 :(
<mnemoc>
today it's just a mess
<mnemoc>
ow
<libv>
-ENODEV :(
<libv>
so otg might be busted on a20
<libv>
but that might also be the script.bin
loulouxiv has quit [Ping timeout: 272 seconds]
loulouxiv has joined #linux-sunxi
<loulouxiv>
hi everyone
bbrezillon has joined #linux-sunxi
SOKS has joined #linux-sunxi
<SOKS>
Hi everyone
<loulouxiv>
yo
<SOKS>
We are trying to use UART0 with a A20 board to print what is typed. Is there any predefined macros that can be used ? Or do we have to do it ourselves using the Datasheet ?
<libv>
SOKS: what do you mean with "print what is typed"?
<oliv3r>
SOKS: erm, i don't think I know what you are trying to do
<oliv3r>
SOKS: e.g. desgin a PCB, talk to the chip via uart
<oliv3r>
SOKS: though at the rate your going now, it's probably best to ask your professor to just flunk you and try a different field
<Wizzup>
Getting two more cubieboard A20s tomorrow \o/
<Wizzup>
Now to get some olimex A20s :)
<oliv3r>
in all fairness, it's called early_printk in u-boot
<oliv3r>
Wizzup: yay :)
<libv>
nah, it's definitely not a bad contact...
<libv>
when it goes, it takes a while until it returns, i haven't figured out which combination brings it back to life, but currently it seems that i need to boot the stock android first
<libv>
there does seem to be something fishy about usb...
<oliv3r>
mripard: re hansg's lradc driver, what If i wanted to use it as a simple adc, i'd have to extend the that driver, to either/also register as some other device (not important) or would there be a 2nd mutually exlusive driver? And how would that work with KConfig, a dual function driver in 2 completly seperated categories?
<libv>
both ohci and ehci disable themselves
<oliv3r>
libv: all 3 properly enabled in the fex?
<oliv3r>
libv: hmm, maybe it's a power issue that causes itself to disconnect?
<libv>
ah, usbc1 is disabled.
<oliv3r>
ahah!
penghb has quit [Read error: Connection timed out]
penghb has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
Black_Horseman has quit [Remote host closed the connection]
<libv>
ok, lsusb now works
<libv>
but it just shows the two root hubs
<libv>
otg still fails and the 8188eu still isn't seen
<libv>
so it doesn't alter anything
<oliv3r>
mripard: I guess it should work the same as his sun4i-ts driver? where you can pass an option to not work as a keyboard, but as an adc
<oliv3r>
libv: let me check something
<oliv3r>
libv: can you show me your fex file?
<libv>
i sent it a few h ago
<libv>
all i altered so far is enabling usbc1
<oliv3r>
ah i saw the u-boot stuff; let me check the fex
<oliv3r>
i'm still backreaeding mail from before xmas
<libv>
:)
<oliv3r>
in reverse order :p
<oliv3r>
or actually random :)
<oliv3r>
libv: don't you have push access to sunxi-boards?
<oliv3r>
mnemoc: ^
<oliv3r>
ah hold off anyway :p
<oliv3r>
libv: while this is about usbc0, strange that it says "axp_ctrl" not seen this before
<hani>
I know that but is there any wiki or info on how to use it with u-boot?
<oliv3r>
hani: there's info on the mailing list, but yes, it's sparse atm
<libv>
oliv3r: you committed the k1001 1gb .fex file btw
<oliv3r>
haha, from the ML :p
<oliv3r>
but if you can modprobe it; then atleast the fex file stuff is working normally
<libv>
no if was initialized, so i think it just silently gave up
wolfy has quit [Ping timeout: 252 seconds]
* libv
waits for the alip image to give up on trying to contact thenetwork again
<libv>
i now set usb_host_init_state = 1 on usbc2
<libv>
as done by the guy who filed the k1001 board
<eagles0513875>
hey all
<wingrime>
nove: ping
<eagles0513875>
oliv3r: going to try the dev branch :)
<eagles0513875>
hopefully that should fix alot of stuff for me :)
<nove>
wingrime, pong
<wingrime>
nove: we have MACC_AVC_CTRL and MACC_AVC_H264_CTRL
<hani>
oliv3r: I've managed to build uboot which is using the mtd nand driver but I'm having trouble booting kernel
<libv>
aha, now the realtek appears
<wingrime>
nove: H264_CTRL better call header
<libv>
modprobe g_ether
<libv>
err, wrong terminal :)
<wingrime>
nove: as AVC_CTRL are for IRQ
<oliv3r>
i fed the troll :(
<oliv3r>
i should work at a zoo
<oliv3r>
sunxi troll feeding
<libv>
he's not a troll, he is a tool
<oliv3r>
hani: i have no experience here, but can you supply a log so we can see?
<wigyori>
morning, happy new year guys
<oliv3r>
wigyori: happy new year to you too
<wingrime>
nove: litte confusing naming
hani has quit [Quit: Page closed]
<libv>
aha, rtl:0x0179 is not known by our 3.4 kernel
<nove>
wingrime, that was what jemk named, i don't know about h264 to know what the register is doing
<wingrime>
nove: are you only player with ISP?
<wingrime>
*played
<oliv3r>
libv: on your a23 tablet, i know you haven't payed much attention, but was the touchscreen driver one of the opensource ones? or only a blob?
<wingrime>
nove: can you send me you repo address
<oliv3r>
libv: easy fix!
<oliv3r>
nove: Jens Kuske
<oliv3r>
nove: oh nvm
<libv>
yeah, google just got me something on that
<mripard>
oliv3r: yes, that's pretty much what I wanted to reply.
<nove>
wingrime, only with jpeg encoding, that used both ISP and AVC
<oliv3r>
mripard: i was not commenting, i was trying to learn ;)
<wingrime>
nove: what are you traced for jpeg
<mripard>
but I got stuck hacking a fish :)
<mripard>
oliv3r: that, plus the fact that using buttons like that is pretty common
<libv>
and i have network...
<mripard>
so it should be generic
<nove>
wingrime, my preliminary jpeg encoding test program is stillnot ready to be public,
<wingrime>
nove: how you figured registers?
<oliv3r>
mripard: i'm not sure i'm following, what I expected though (and this is my lack of knowledge again) is a plain adc driver, with some 'common' framework that can take an adc as input, and output a keyboard/input thing
<oliv3r>
libv: *high five*
<wingrime>
nove: no,no I about your trace analyser
<libv>
and wikied :)
<libv>
still, otg is a mystery still
<oliv3r>
libv: and patched?
<nove>
wingrime, i used libjpegenc.a from android libs, is only a 1 function api
<eagles0513875>
btw guys happy new year :)
<oliv3r>
libv: well it looked like stuff was missing there too?
<libv>
oliv3r: this is one of the few tablets that has no separate power connector
<libv>
it charges solely over usb
<libv>
this might be part of things...
<wingrime>
nove: and it worked?
<mripard>
oliv3r: yes, and the keyboard/input thing can be generic
<oliv3r>
libv: hmm, it oculd be they castrated things a little; but in theory, it shouldn't matter how it's powerd
<mripard>
mripard: it's exactly what I had in mind :)
<oliv3r>
mripard: but this doesn't exist yet?
<mripard>
for such buttons, I don't think so
<libv>
yea...
<oliv3r>
mripard: that supprises me actually
<mripard>
the ADC framework is quite new
pacopad_ has joined #linux-sunxi
<nove>
wingrime, yes android .a libs somewhat works, after adding some stubs function for missing symbols
<oliv3r>
mripard: i see; challange hansg! he has some spare time these next few days!
<wingrime>
nove: nice work
<wingrime>
nove: also, you would be helpfull with ac3 and dts
<oliv3r>
libv: i'm not sure what the restrict_flag does; usbc0 section seems to have grown bigger then last time i saw it; but the id_gpio is required
<wingrime>
nove: we also _have_ only android sources
<oliv3r>
libv: usb_port_type = 2 says 'both' iirc, so it needs the ID gpio to know what it really is, as that gpio shorts to gnd iirc
<wingrime>
nove: not sure it present in a20 and next SoCs
<wingrime>
libv: can you try vdpau with a23 some time?
<libv>
wingrime: sure...
<libv>
wingrime: when it runs on sunxi u-boot and sunxi kernel
<wingrime>
libv: you get aw's uboot?
SOKS has quit [Ping timeout: 252 seconds]
loulouxiv has quit [Read error: Operation timed out]
<libv>
wingrime: i have put the a23 tablet aside
neokeld has quit [Read error: Operation timed out]
<wingrime>
libv: so, what are you finaly get with a23?
<libv>
these things are like 40EUR, i am surprised that noone else has them yet :)
<nove>
wingrime, yesterday, i tried libjpegenc.a from A23 sdk, has the same problems as the one from A31 sdk, there are hardware differences, its writes 000300c7 to MACC_VE_CTRL
<libv>
wingrime: i will work a23 when i have finished my kms driver and my lima driver
<wingrime>
libv: kms ?
<wingrime>
witch one
<libv>
...
<libv>
sunxi?
<wingrime>
sunxi or lima?
<libv>
oliv3r: i really have a feeling that our sunxi-3.4 kernel is leaving some bits out from the stock android kernel...
<libv>
and i am pretty worried about the serial issues i sometimes see
pacopad has quit [Read error: Operation timed out]
<wingrime>
nove: not sure HW realy diffrent
<wingrime>
nove: only 0xc0 are different
<wingrime>
nove: It can be some flag or something unknown for me
<wingrime>
0x030007
<wingrime>
are realy normal for init process
<oliv3r>
libv: but why hasn't anybody commented on it before?
<wingrime>
nove: you can try add "c" for jemks JPEG sample or vdpau
<oliv3r>
libv: though it surely is possible mnemoc is responsible for those merges i think
<oliv3r>
mnemoc: ^
pacopad has joined #linux-sunxi
<nove>
wingrime, after its set that, it starts setting the registers in the same way as the older version (that works in A13), that failed because the subengine is not active
<wingrime>
nove: thats interesting
<nove>
wingrime, i think the that the subengine is enabled changed with A31/A23
<nove>
*the way
<wingrime>
nove: not sure they will chane that without a reason
<wingrime>
*change
pacopad_ has joined #linux-sunxi
<wingrime>
nove: but
<wingrime>
nove: c7 setted up just after jpeg INIT?
<eagles0513875>
hey guys if i stop using buildroot what does one need to do to just get a sunxi system going on an embedded device such as my cubieboard 2
<oliv3r>
ssvb: ping
<wingrime>
nove: and also, IP version not checked by blob?
<eagles0513875>
oliv3r: do you know if i use the revision for the latest dev branch if in need to bump my kernel version?
<nove>
wingrime, i will get you a trace
<oliv3r>
eagles0513875: i don't use buildroot; so no clue
pacopad has quit [Ping timeout: 252 seconds]
<wingrime>
nove: can you post link to yours trace tool repo
<nove>
wingrime, libjpegenc.a doesn't check for version
<eagles0513875>
ok if i want to drop using build root what would i need to do to get an embedded system working with the bare essentials and a root system only
<wingrime>
nove: thats can be also means, that they added new engines???
neokeld has joined #linux-sunxi
pacopad_ has quit [Ping timeout: 246 seconds]
<nove>
wingrime, post link of what? you want the binary blob jpeg encoder to do tracing?
<libv>
oliv3r: well, i am one of the few people who does seem to have bought and tested an a20 tablet
pacopad has joined #linux-sunxi
vector80 has joined #linux-sunxi
<libv>
apart from the k1001, the two other a20 definitely will not have working wifi
<wingrime>
nove: no, your trace tool (visualizer)
<vector80>
Hi, did anybody try streaming camera on A20 ?
<libv>
i also doubt that anyone tried OTG on a20 with our kernel
<libv>
n01: apparently, you are called carlo caione, and i found that out since you mentioned this on the mainlining page
<libv>
n01: i have now changed the account carlo to "people"
<libv>
n01: do put your realname up in your irc client
<n01>
libv: yep, that's my real name. I'll put it on irc
<libv>
n01: /whois n01 is really not helpful as it is
<wingrime>
ae0 and ae4 looks like SRAM regs
kivutar has quit [Read error: Operation timed out]
pacopad has joined #linux-sunxi
<nove>
wingrime: it is, a30, a34 appears to be is the output image x, y offset for luma and chroma, a38 is related to amount of scaling, but this needs to be more tested
<wingrime>
nove: I added ae0 and ae4 to wiki
sspiff has quit [Quit: Konversation terminated!]
pacopad_ has joined #linux-sunxi
<wingrime>
nove: but you need check a20/a10 presence
<wens>
Turl: curious that sun4i~6i dts still have dummy clock for 200M output from PLL6 to cpu
<wingrime>
wens: HDMI IP block have own PLL or two
<wens>
sun7i uses <&pll6 1>, but that is at 600M, which is wrong
<wens>
wingrime: I'm doing rename of clock nodes in dts, not adding new ones atm
<wingrime>
wens: better add IRQ names...
<wingrime>
nove: MACC_ISP_WB_THUMB_ADDR_C , Y it's from blob, but I need check it, may be I got wrong address
<wingrime>
looks it shifter to one doubleword
<wingrime>
*shifted
pacopad has quit [Ping timeout: 260 seconds]
<wingrime>
nove: WriteBack Thumb addr chroma
pacopad has joined #linux-sunxi
ZetaNeta has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
hipboi has quit [Remote host closed the connection]
<wens>
mripard: we have 4 clocks (cpu, apb0, ahb, axi) at the same address, clk@01c20054
<wens>
any suggestions? leave them be?
pacopad_ has quit [Ping timeout: 260 seconds]
<mripard>
wens: yeah, let them like this
<wens>
leave them as they originally were? {cpu,apb0,ahb,axi}@01c20054?
<mripard>
yep
<wens>
ok
<wens>
I will keep the original names for dummy clocks, as they will either be removed or renamed in the future
loulouxiv has joined #linux-sunxi
EGM has quit [Remote host closed the connection]
n01 has quit [Ping timeout: 257 seconds]
<andhe>
wens: if I've understood it correctly, the bluetooth parts in your wifi branch isn't expected to work yet?
<wens>
andhe: that's right
<andhe>
wens: have you done any investigations yet or have any clue as to why?
<wens>
i tried all i could think of
<andhe>
ok. just checking.
<wens>
mine doesn't work on 3.4 kernel either, so it could be broken
<wens>
anyway, bluetooth shouldn't require any driver changes, just adding the relevant bits to the dt
<andhe>
I'm a noob, so just trying to wrap my head around all of the devicetree bits and comparing to the hw manual... still has quite a bit to find out to fully understand even simple stuff like how all of the uart2 gets configured...
<andhe>
bluetooth seemed like a simple enough thing to start looking at.... AIUI just get the uart up and then it should be time for the userspace configure program to send it commands to get bt chip to reconfigure itself in h4 mode.... and the rfkill parts on top of that....
<wens>
it doesn't help that we don't have the manual for AP6210 , so we don't know how the external pins map to the ICs
<andhe>
but I should start with confirming I can get it to work on 3.4 first.... still haven't done that. :P
<wens>
you could try booting android off nand first, and see if that works
<wens>
I haven't done it on mine yet, though
<andhe>
good idea
<andhe>
haven't really looked much at android at all yet.... but it's still sitting there on the nand.
<wens>
i'd like to get the rfkill and brcmfmac bits merged first, if possible
<andhe>
(I just got my cubietruck and have just started to get up to speed with linux-sunxi project and the mainlining effort.)
_massi_ has quit [Quit: Leaving]
<andhe>
wens: good idea... wifi seems to work here..... (only tried scanning for ssids though)
bbrezillon has quit [Quit: Ex-Chat]
pacopad has joined #linux-sunxi
FR^2 has quit [Quit: Connection reset by peer]
<andhe>
I think I've grasped how the pin muxing options gets specified, but not yet located where to see the uart2 pin direction in/out configuration and how the driver knows which pin is RTS/CTS/RX/TX... (hmmm... is it implicitly specified by the number of pins and their order in the devicetree?)
enrico_ has quit [Quit: Bye]
pacopad has quit [Ping timeout: 246 seconds]
Gerwin_J has joined #linux-sunxi
pwhalen has quit [Ping timeout: 272 seconds]
Quarx has quit []
libv has quit [Ping timeout: 260 seconds]
SOKS has quit [Quit: SOKS]
neokeld has quit [Quit: leaving]
panda84kde has quit [Quit: Konversation terminated!]
loulouxiv has quit [Ping timeout: 272 seconds]
pwhalen has joined #linux-sunxi
pacopad has joined #linux-sunxi
gzamboni has quit [Ping timeout: 252 seconds]
pacopad has quit [Ping timeout: 260 seconds]
gzamboni has joined #linux-sunxi
rellla has quit [Ping timeout: 260 seconds]
<oliv3r>
lioka: that's actual a fair port
<oliv3r>
vector80: that you used the wrong bootloader :p
<oliv3r>
vector80: you are probably using the mmc bootloader on the emmc without a valid u-boot on mmc
<libv>
oliv3r: the current status page there is quite interesting
<libv>
"The current linux-sunxi (as of Nov 21 2013) seems to have broken USB power on many devices with default defconfig settings (sun7i_defconfig). With the sunxi-3.4 branch, this can be fixed by enabling the Inventra Dual Role USB 2.0 Controller driver with the Allwinner platform glue option. With the stage/sunxi-3.4 branch even that doesn´t help."
<oliv3r>
hmm, that is strange
<oliv3r>
hglm ping!
<wolfy>
oliv3r: incorrect usage. should be ping hglm :)
notmart has joined #linux-sunxi
lgentili has joined #linux-sunxi
NightShade has joined #linux-sunxi
lgentili has quit [Quit: Page closed]
<libv>
oliv3r: you have a proper Q88, right?
<libv>
a13 tablet?
cajg has quit [Ping timeout: 240 seconds]
<libv>
ah, 713b
brain_ has joined #linux-sunxi
t3st3r has quit [Remote host closed the connection]
hramrach has quit [Remote host closed the connection]
<brain_>
Can anyone help me get a ft5x_ts driver to properly initialize? i've been trying everything. im on 3.0.8 and no way to change it due to some driver issues for android.
<brain_>
it seems the only error I get is "ctp_init_platform_resource: tp_wakeup request gpio fail!"
t3st3r has joined #linux-sunxi
<brain_>
i think it has to go with the missing gpio driver but i could be wrong, it works fine with the stock driver
hramrach has joined #linux-sunxi
<brain_>
i've done every tweak i could think of to the code but i get no output when i cat /input...getevent shows the device registered but does not respond to events
<brain_>
it could also be an interupt issue but i dont know how to debug that
cajg has joined #linux-sunxi
<Turl>
brain_: did you check your fex config?
<brain_>
fex config is fine, it works with the stock driver, wierd thing tho is i2cdetect shows the panel on 0x38 but the fex and oem driver detect and register on address 0x70. i did try changing the fex to 0x38 but got the same results. the stock driver i reversed using ida and got alot of info but nothing i tweak makes it show output.
<brain_>
i am wondering if their is a way of changing were the driver looks for the device but i dont see anything other than the fex config. also if it would help maybe i could try booting linux from an SD card and using some of those tools to further debug things? im just out of ideas.
<brain_>
i2cdump shows the device responding to touch fine on all the registers
<mrnuke>
oliv3r: I have a patch to do udelay in raminit. github pull request I assume?
<netlynx>
brain_, those addresses 0x38 (00111000b) and 0x70 (01110000b) differs on a power of 2. Are FEX and oem driver using a numbering scheme that includes the R/W bit in the address?
<brain_>
that i'll check into
gzamboni_ has quit [Ping timeout: 246 seconds]
RichiH has left #linux-sunxi [#linux-sunxi]
AreaScout has quit [Ping timeout: 265 seconds]
pacopad_ has joined #linux-sunxi
<eagles0513875>
hey guys what is the difference between uimage and zimage
kivutar has joined #linux-sunxi
<NightShade>
uImage is packed ready for u-boot to load
<NightShade>
zImage is a compressed kernel, that is then packed into uImage
<eagles0513875>
ok then i have a problem for some reason when i generate my build root setup for my device generating a zimage
<NightShade>
have you got the u-boot tools installed?
gzamboni_ has joined #linux-sunxi
<pacopad_>
Hi Guys i can't launch gles2 app on my cb2, i installed toubo-fb and mali drives but nothing
<pacopad_>
always error messages
<pacopad_>
what must i check ?
<eagles0513875>
NightShade: im using build root and it takes care of what i need installed
<eagles0513875>
but im wondering how hard it would be if i were to dump using build root and build my own root system from scratch
gzamboni_ has quit [Ping timeout: 264 seconds]
<hramrach>
eagles0513875 see the wiki for build instructions
<eagles0513875>
ok. still not sure whta im going to do yet
<hramrach>
also you need something like u-boot-tools to generate uImage
gzamboni_ has joined #linux-sunxi
<hramrach>
without those you get zImage and the build stops there
<hramrach>
mrnuke: preferably send to the mailing list. Especially since it's a single patch
<mrnuke>
hramrach: I hate ML patches. Is this really needed? :(
<n01>
mrnuke: yes :)
<hramrach>
well, the maintainers of the trees like them ..
<n01>
so people con review it
<n01>
s/con/can
<hramrach>
pacopad_: the test program that came with mali
<pacopad_>
it returns an error
<hramrach>
then you installed mali wrong
<pacopad_>
Eglinitialsie failed
<hramrach>
any other info?
<pacopad_>
nope
<pacopad_>
i just did make and make install
<pacopad_>
what could be wrong :)
<hramrach>
do the install again and make sure you select X11 libs
<pacopad_>
i check that
<pacopad_>
the compiled lib is in the x11 folder
<pacopad_>
of the mali source driver
<hramrach>
there are compiled libs in every folder. It's binary-only driver
<pacopad_>
:)
<pacopad_>
could the auto detect part of the make return bad info about mali version ?
<hramrach>
it should not but it can't select fb or x11 version automagically
<eagles0513875>
then there is something a bit odd
<eagles0513875>
with build root
<eagles0513875>
default it is set to generate uimage yet this time around it generated a zimage O_o
<hramrach>
it must generate a zimage to turn it into uimage
<eagles0513875>
hramrach: is there a performance difference between zimage and uimage
<hramrach>
it's a different image format of the same kernel
<hramrach>
pacopad_: that ? may not work
<pacopad_>
it's the script output
quintin_ has quit [Quit: Leaving.]
<hramrach>
pacopad_: you can always set that by hand so do it
<pacopad_>
same :(
<hramrach>
never tried it on cb2 tbh
gzamboni_ has quit [Ping timeout: 240 seconds]
<NightShade>
eagles0513875: what does the command mkimage say on your build machine?
<libv>
hrm... and there the serial goes again
<libv>
but after the rtl gave up
<eagles0513875>
ok im getting an error now im using the sunxi devel branch so i think i just need to switch the setting in build root to use zimage instead NightShade
<libv>
something more fundamental is going wrong with this hw
<NightShade>
you need to install the tool that converts from zImage to a uImage it's not normally installed by default
<NightShade>
it's in the package uboot-tools in fedora or u-boot-tools in debian
<eagles0513875>
ok
pacopad_ has quit [Quit: pacopad_]
n01 has quit [Ping timeout: 240 seconds]
<mrnuke>
hramrach: haha! it ended up being two patches
ZetaNeta has quit [Ping timeout: 272 seconds]
<mrnuke>
hramrach: so it stays on github
pacopad has quit [Quit: pacopad]
ZetaNeta has joined #linux-sunxi
wingrime has quit [Ping timeout: 252 seconds]
<libv>
hrm, hramrach, the mali-libs README explicitely states how to force certain config value
<libv>
ah, no, it's still up... it apparently is some networking idiocy, and me having put a huge usb extension cable on
<eagles0513875>
NightShade: i installed the tools yet it still errors
<eagles0513875>
will there be any performance difference between uimage and zimage?
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<NightShade>
u-boot can't load a zImage directly as I understand it
<eagles0513875>
no no what im asking are there performance differences NightShade
<NightShade>
none at all
<eagles0513875>
ok
<NightShade>
it just won't work
<eagles0513875>
?
<NightShade>
the u-boot boot loader doesn't know how to load and run a zImage kernel
<NightShade>
so while it's the same kernel it has none of the packaging that allows uboot to understand it
<NightShade>
for example this command: mkimage -A arm -O linux -T kernel -C none -n uImage -a 0x00008000 -e 0x00008000 -d vmlinuz uImage
<NightShade>
that packages a zImage kernel up into a uImage that uboot understands, but thta's for a very different arm board.
<NightShade>
notice it's telling uboot where in memory to put the kernel etc
NightShade has quit [Ping timeout: 260 seconds]
awafaa has quit [Ping timeout: 246 seconds]
bsdfox has quit [Ping timeout: 245 seconds]
gzamboni has joined #linux-sunxi
notmart has quit [Quit: notmart terminated!]
nove has quit [Quit: nove]
wingrime has joined #linux-sunxi
bsdfox has joined #linux-sunxi
steev has quit [Ping timeout: 252 seconds]
awafaa has joined #linux-sunxi
gzamboni has quit [Ping timeout: 246 seconds]
steev has joined #linux-sunxi
gzamboni has joined #linux-sunxi
ojn has quit [Ping timeout: 246 seconds]
steev has quit [Ping timeout: 240 seconds]
pwhalen has quit [Ping timeout: 246 seconds]
gzamboni has quit [Ping timeout: 264 seconds]
gzamboni has joined #linux-sunxi
pwhalen has joined #linux-sunxi
pwhalen has joined #linux-sunxi
pwhalen has quit [Changing host]
ganbold_ has quit [Remote host closed the connection]